Stats
Actions
Tags
From qe-framework
Extracts factual claims from documents and verifies them through evidence-based research, producing claim-by-claim ratings.
How this skill is triggered — by the user, by Claude, or both
Slash command
/qe-framework:Qfact-checkerThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Extract factual claims from documents, reports, and articles and systematically verify them.
Extract factual claims from documents, reports, and articles and systematically verify them.
1. Claim extraction → 2. Prioritization → 3. Evidence collection → 4. Source verification → 5. Verdict → 6. Report writing
| Target | Example |
|---|---|
| Factual claims | "X happened", "Y is true" |
| Statistics/numbers | "40% increase", "1 million people" |
| Dates/times | "Started in January 2024" |
| Citations/attribution | "A said..." |
| Causal claims | "X caused Y to happen" |
## Claim Log
**Document:** [Source]
**Date:** [Write date]
### Claim 1
- **Statement:** [Exact quote or summary]
- **Speaker:** [Who said it]
- **Context:** [Surrounding context]
- **Type:** Statistic / Historical fact / Citation / Causal
- **Priority:** High / Medium / Low
- **Status:** Unverified / Verified / False / Unverifiable
| Priority | Criteria |
|---|---|
| High | Core thesis of document, easily verifiable, high impact if wrong |
| Medium | Supporting evidence, requires effort to verify |
| Low | Peripheral details, commonly accepted, minimal impact |
| Claim Type | Primary Source |
|---|---|
| Statistics | Original research, government data, methodology |
| Citations | Audio/video recordings, transcripts, direct verification |
| History | Contemporary news reports, official records |
| Science | Peer-reviewed papers, expert consensus |
| Legal | Court documents, official filings |
| Financial | Audit reports, regulatory disclosures |
| Evidence Type | Strength |
|---|---|
| Official documents (court records, government reports) | Strong |
| Original data (directly analyzable) | Strong |
| Expert consensus (multiple independent experts) | Strong |
| Named sources (direct knowledge) | Medium |
| Contemporary reporting (same-era news) | Medium |
| Anonymous sources | Weak |
| Social media posts | Weak |
| Grade | Criteria |
|---|---|
| True | Accurate and complete, no important omissions |
| Mostly True | Accurate but needs context or minor clarification |
| Half True | Partially accurate, key context missing |
| Mostly False | Some factual content but overall misleading |
| False | Contradicts evidence |
| Unverifiable | Insufficient evidence |
## Verdict: [Claim Summary]
**Claim:** [Exact statement]
**Speaker:** [Who]
**Verdict:** [Grade]
### Supporting Evidence
- [Evidence 1 + Source]
- [Evidence 2 + Source]
### Contradicting Evidence
- [Evidence 1 + Source]
### Missing Context
- [Context 1]
### Verdict Reasoning
[Why this grade]
### Confidence
High / Medium / Low + Reason
# Fact-Check Report
**Target Document:** [Title/URL]
**Verification Date:** [Date]
**Verifier:** [Name]
## Summary
- Total claims: N
- True: N | Mostly true: N | False: N | Unverifiable: N
## Key Findings
### [Claim 1] — Verdict: [Grade]
- **Original:** "[Quote]"
- **Evidence:** [Summary of basis]
- **Correction suggestion:** [If needed]
### [Claim 2] — Verdict: [Grade]
...
## Corrections Needed
| Location | Original | Correction | Reason |
|----------|----------|------------|--------|
| [Section/Line] | [Original] | [Correction] | [Basis] |
## Sources
1. [Source 1]
2. [Source 2]
npx claudepluginhub inho-team/qe-framework --plugin qe-frameworkCreates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.